Update app.py

app.py

@@ -8,6 +8,27 @@ os.environ['OPENAI_API_KEY'] = os.environ.get('OPENAI_API_KEY')
 
 client = OpenAI() # add api_key
 
+import torch
+import torchaudio
+import gradio as gr
+from scipy.io import wavfile
+from scipy.io.wavfile import write
+
+knn_vc = torch.hub.load('bshall/knn-vc', 'knn_vc', prematched=True, trust_repo=True, pretrained=True, device='cpu')
+
+def voice_change(audio_in, audio_ref):
+    samplerate1, data1 = wavfile.read(audio_in)
+    samplerate2, data2 = wavfile.read(audio_ref)
+    write("./audio_in.wav", samplerate1, data1)
+    write("./audio_ref.wav", samplerate2, data2)
+
+    query_seq = knn_vc.get_features("./audio_in.wav")
+    matching_set = knn_vc.get_matching_set(["./audio_ref.wav"])
+    out_wav = knn_vc.match(query_seq, matching_set, topk=4)
+    torchaudio.save('output.wav', out_wav[None], 16000)
+    return 'output.wav'
+
+
 def tts(text, model, voice):
     response = client.audio.speech.create(
         model=model, #"tts-1","tts-1-hd"
@@ -25,18 +46,34 @@ def tts(text, model, voice):
     return temp_file_path
 
 
-with gr.Blocks() as demo:
-    gr.Markdown("# <center> OpenAI Text-To-Speech API with Gradio </center>")
-    gr.HTML("You can also access the Streaming demo for OpenAI TTS by clicking this <a href='https://huggingface.co/spaces/ysharma/OpenAI_TTS_Streaming'>Gradio demo link</a>")
+app = gr.Blocks()
+
+with app:
+    gr.Markdown("# <center>🥳🎶🎡 - KNN-VC AI变声</center>")
+    gr.Markdown("### <center>🌟 - 3秒实时AI变声，支持中日英在内的所有语言！无需训练、一键变声！🍻 </center>")
+    gr.Markdown("### <center>🌊 - 更多精彩应用，敬请关注[滔滔AI](http://www.talktalkai.com)；滔滔AI，为爱滔滔！💕</center>")
+
     with gr.Row():
-      model = gr.Dropdown(choices=['tts-1','tts-1-hd'], label='Model', value='tts-1')
-      voice = gr.Dropdown(choices=['alloy', 'echo', 'fable', 'onyx', 'nova', 'shimmer'], label='Voice Options', value='alloy')
+      with gr.Column():
+        inp_text = gr.Textbox(label="请填写您想生成的文本（中英文皆可）", placeholder="想说却还没说的 还很多 攒着是因为想写成歌")
+        btn_text = gr.Button("一键开启真实拟声吧", variant="primary")
+
+      with gr.Column():
+        inp1 = gr.Audio(type="filepath", label="请上传AI变声的原音频（决定变声后的语音内容）")
+        inp2 = gr.Audio(type="filepath", label="请上传AI变声的参照音频（决定变声后的语音音色）")
+        btn1 = gr.Button("一键开启AI变声吧", variant="primary")
+      with gr.Column():
+        out1 = gr.Audio(type="filepath", label="AI变声后的专属音频")
+      btn_text.click(tts, inp_text, inp1)
+      btn1.click(voice_change, [inp1, inp2], out1)
+
+    gr.Markdown("### <center>注意❗：请不要生成会对个人以及组织造成侵害的内容，此程序仅供科研、学习及个人娱乐使用。</center>")
+    gr.HTML('''
+        <div class="footer">
+                    <p>🌊🏞️🎶 - 江水东流急，滔滔无尽声。 明·顾璘
+                    </p>
+        </div>
+    ''')
 
-    text = gr.Textbox(label="Input text", placeholder="Input text and press the Text-To-Speech button or press Enter.")
-    btn = gr.Button("Text-To-Speech")
-    output_audio = gr.Audio(label="Speech Output")
-    
-    text.submit(fn=tts, inputs=[text, model, voice], outputs=output_audio, api_name="tts", concurrency_limit=None)
-    btn.click(fn=tts, inputs=[text, model, voice], outputs=output_audio, api_name="tts", concurrency_limit=None)
+app.launch(show_error=True)
 
-demo.launch()